# Replication Package: Cross-Paper Fragility and Robustness Assessment

## Overview
This folder contains all files needed to replicate the analysis in "Cross-Paper Fragility and Robustness Assessment." This meta-analytical paper examines the robustness of demographic effects on capital flows across all papers in the project. It includes sample composition statistics, unified scorecards, cross-paper jackknife tests, influential country identification, and multiple testing corrections.

## Requirements
- Python 3.10+
- pandas, numpy, scipy, statsmodels
- Data files in data/processed/ (note: fragility scripts may read from other project directories)

## Structure
- `scripts/` — Analysis scripts (run in phase order)
- `src/` — Shared modules (PanelGLS estimator, data loading, country classifications)
- `data/processed/` — Processed panel data (if any)
- `output/tables/` — Generated output tables
- `paper/` — Paper manuscript and references

## Reproduction
Run scripts in numerical phase order:
```
python scripts/phase1_composition_statistics.py
python scripts/phase2_unified_scorecard.py
python scripts/phase3_cross_paper_jackknife.py
python scripts/phase4_influential_countries.py
python scripts/phase5_multiple_testing.py
```

## Data Sources
- UN World Population Prospects 2024
- IMF World Economic Outlook
- Penn World Table 10.01
- Chinn-Ito KAOPEN Index
- Lane & Milesi-Ferretti External Wealth of Nations
- Output from all other project papers

## Notes
- All analysis uses the 140-country expanded panel (EBA-49 + SSA-20 + EU expansion + Tier 1 expansion)
- The `src/` modules are from the multilateral/followup project and contain the expanded country lists
- This paper reads results/data from other project directories; paths may need adjustment
- Multiple testing corrections applied across all project results
